System and method for signature-based systematic condition detection and analysis

• US 7,853,848 B2
• Filed: 10/22/2007
• Issued: 12/14/2010
• Est. Priority Date: 10/22/2007
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kassab, Maroun, Desineni, Rao H., Pastel, Leah M.
• Primary Examiner(s): Tabone, Jr.; John J
• Application Number:
  US11/875,975
• Publication Number:
  US 20090106614A1
• Time in Patent Office: 1,149 Days
• Field of Search: 714/724, 714/732, 714/742, 714/745, 714/E11.159, 324/528, 324/531, 324/765, 716/4, 438/14, 438/17
• US Class Current: 714/732
• CPC Class Codes: G01R 31/318321 for combinational circuits

Abstract

Disclosed are embodiments of a system, method and service for detecting and analyzing systematic conditions occurring in manufactured devices. Each embodiment comprises generating a unique signature for each of multiple tested devices. The signatures are generated based on an initial set of signature definitions and the values for those signature definitions that are derived at least in part from selected testing data. A systematic condition is detected based on commonalities between the signatures. The systematic condition is then analyzed, alone or in conjunction with additional information, in order to develop a list of underlying similarities between the devices. The analysis results can be used to refine the systematic condition detection and analysis processes by revising the signature definitions set and/or by modifying data selection.

Claims

1. A system for detection and analysis of systematic conditions occurring in manufactured devices, said manufactured devices being manufactured according to same design and ground rules and said system comprising:

a signature generator receiving testing results from essentially identical and repeatable testing performed on said manufactured devices and further generating, based on said testing results, corresponding signatures for said manufactured devices each signature comprising values for an initial set of signature definitions and said signature definitions comprising at least one of electrical properties, physical properties, performance information, environmental information, statistically abnormal data, power metrics, electrically coded information and manufacturing process information;

a systematic condition detector in communication with said signature generator and detecting a systematic condition occurring in at least some of said manufactured devices based on commonalities between at least some of said corresponding signatures for said manufactured devices, said systematic condition comprising a condition that is observed and that occurs non-randomly;

a systematic condition analyzer in communication with said systematic condition detector and analyzing said systematic condition by analyzing said at least some of said corresponding signatures having said commonalities to develop a list of underlying similarities between said at least some of said manufactured devices, said list comprising any of characteristics, features, parameters, defects and defect mechanisms indicative of at least one potential root cause of said systematic condition;

a signature definitions selector; and

a data selector,

said systematic condition analyzer further outputting analysis results to at least one of said signature definitions selector and said data selector,

said signature definitions selector further automatically creating a revised set of said signature definitions, based on said analysis results,

said data selector further automatically selecting different testing data, based on said analysis results, and

said signature generator further generating new signatures using at least one of said revised set and said different testing data.

2. The system of claim 1, said systematic condition detector further detecting said systematic condition in other devices, based on said new signatures.

3. The system of claim 1, said systematic condition analyzer further amending said list, based on at least one of said new signatures and other known signatures.

4. The system of claim 1, said systematic conditions analyzer further analyzing said systematic condition, based on additional information.

5. The system of claim 4, said additional information comprising at least one of known root causes of known signatures, design information, manufacturing process information, performance information, and diagnostic testing information.

6. The system of claim 1, signature definition selector further automatically creating said revised set by one of modifying said initial set and creating a completely new set.

7. The system of claim 1, said systematic condition analyzer further automatically outputting said analysis results to at least one of designers, manufacturers, testers and yield analyzers so that at least one of ground rules, design specifications and layout, manufacturing processes, testing, datalogging and yield analysis may be updated.

8. A method for detection and analysis of systematic conditions occurring in manufactured devices, said manufactured devices being manufactured according to same design and ground rules and said method comprising:

receiving testing results from essentially identical and repeatable testing performed on said manufactured devices;

generating, based on said testing results, corresponding signatures for said manufactured devices, each signature comprising values for an initial set of signature definitions and said signature definitions comprising at least one of electrical properties, physical properties, performance information, environmental information, statistically abnormal data, power metrics, electrically coded information and manufacturing process information;

detecting a systematic condition occurring in at least some of said manufactured devices based on commonalities between at least some of said corresponding signatures for said manufactured devices, said systematic condition comprising a condition that is observed and that occurs non-randomly;

analyzing said systematic condition by analyzing said at least some of said corresponding signatures having said commonalities to develop a list of underlying similarities between said at least some of said manufactured devices, said list comprising any of characteristics, features, parameters, defects and defect mechanisms indicative of at least one potential root cause of said systematic condition;

at least one of creating a revised set of signature definitions, based on analysis results, and selecting different testing data, based on said analysis results; and

generating new signatures using at least one of said revised set and said different testing data.

9. The method of claim 8, further comprising at least one of detecting said systematic condition in other devices based on said new signatures;

and, amending said list based on at least one of said new signatures and other known signatures.

10. The method of claim 8, said analyzing of said systematic condition further comprising analyzing said systematic condition based on additional information.

11. The method of claim 10, said additional information comprising at least one of known root causes of known signatures, design information, manufacturing process information, performance information, and diagnostic testing information.

12. The method of claim 8, said creating of said revised set of signature definitions comprising one of modifying said initial set of signature definitions and creating a completely new set of signature definitions.

13. The method of claim 8, further comprising updating at least one of ground rules, design specifications and layout, manufacturing processes, testing, yield analysis and datalogging, based on said analysis results.

14. The method of claim 8, further comprising at least one of binning and dispositioning said at least some of said manufactured devices, based on said systematic condition.

15. A service for detection and analysis of systematic conditions occurring in manufactured devices, said manufactured devices being manufactured according to same design and ground rules and said service comprising:

receiving testing results from essentially identical and repeatable testing performed on said manufactured devices;

generating, based on said testing results, corresponding signatures for said manufactured devices, each signature comprising values for an initial set of signature definitions and signature definitions comprising at least one of electrical properties, physical properties, performance information, environmental information, statistically abnormal data, power metrics, electrically coded information and manufacturing process information;

detecting a systematic condition occurring in at least some of said manufactured devices based on commonalities between at least some of said corresponding signatures for said manufactured devices, said systematic condition comprising a condition that is observed and that occurs non-randomly;

analyzing said systematic condition by analyzing said at least some of said corresponding signatures having said commonalities to develop a list of underlying similarities between said at least some of said manufactured devices, said list comprising any of characteristics, features, parameters, defects and defect mechanisms indicative of at least one potential root cause of said systematic condition;

at least one of creating a revised set of signature definitions, based on analysis results, and selecting different testing data, based on said analysis results; and

generating new signatures using at least one of said revised set and said different testing data.

16. The service of claim 15, said analyzing of said systematiac condition further comprising analyzing said systematic condition based on additional information.

17. The service of claim 16, said additional information comprising at least one of known root causes of known signatures, design information, manufacturing process information, performance information, and diagnostic testing information.

18. A computer program product comprising a computer-readable storage medium, said storage medium readable by a computer and storing a program of instructions executable by said computer to perform a method for detection and analysis of systematic conditions occurring in manufactured devices, said manufactured devices being manufactured according to same design and ground rules and said method comprising:

receiving testing results from essentially identical and repeatable testing performed on said manufactured devices;

generating, based on said testing results, corresponding signatures for said manufactured devices, each signature comprising values for an initial set of signature definitions and said signature definitions comprising at least one of electrical properties, physical properties, performance information, environmental information, statistically abnormal data, power metrics, electrically coded information and manufacturing process information;

detecting a systematic condition occurring in at least some of said manufactured devices based on commonalities between at least some of said corresponding signatures for said manufactured devices, said systematic condition comprising a condition that is observed and that occurs non-randomly;

analyzing said systematic condition by analyzing said at least some of said corresponding signatures having said commonalities to develop a list of underlying similarities between said at least some of said manufactured devices, said list comprising any of characteristics, features, parameters, defects and defect mechanisms indicative of at least one potential root cause of said systematic condition;

at least one of creating a revised set of signature definitions, based on analysis results, and selecting different testing data, based on said analysis results; and

generating new signatures using at least one of said revised set and said different testing data.